Calibration for rice yield model

[kgbrusola]

Hi everyone!

I am new to APSIM and trying to hindcast and simulate rice yield. I already have the yield data and I was wondering how do I calibrate the model first? I have tried and understood the examples from the website. Thank you.

[peter-devoil]

@andomariot and @DonaldGaydon may have better pointers here; but in general:

• get your simulation(s) in shape to match your experiment, use linked components wherever possible across treatments
• check that the simulation runs with your met record and soils
• Phenology: ensure you have flowering and maturity correct (may involve creating a new cultivar )
• Biomass: examine biomass and water uptake in conjuction
• 9 times out of 10; the correct yield will be looking at you now once you've verified the above

[kgbrusola]

Hi @peter-devoil!

Thank you for your comment. May I know what do the photop-sens and vern-sens variables mean? I have not explored other cultivars used in APSIM.

Thank you so much.

[peter-devoil]

Oryza's photoperiod sensitivity (PPSE) is discussed on page 34.

I don't think it has vernalisation - that's a wheat thing.

[kgbrusola]

Thanks for the reference. I am currently reading it and studying which value to use depending on the variety of the crop planted in my ROI. I have another question. My study area reported 4 values of yield which is every 15 days. For example, for September 1-15 yield is 6 tons/ha then September 16-30 yield is 6.5 tons/ha and so on. I have observed that the result of the simulated yield only have 45 days duration. How can I make it 60 days or should I only get the highest yield per 15 days? Thanks.

[andomariot]

Hi, Pete comments are spot on for the calibration steps. I am not sure to understand though your above questions on number of days with different yield values. Are these windows of sowing dates? Or dates when you have observation of yield from experiments? If these are observed data, it is very useful for the calibration as you can get the model to capture the trends of biomass production from your experiments, the more data you have the better your calibration and your accuracy. You better use then all the values. I hope this helps

[kgbrusola]

Hi @andomariot. The dates that I mentioned were the time the actual yield in the area were reported. Meanwhile, the 45 days that I mentioned was from APSIM where there is a yield value on the results. Can I input the observed data in APSIM and will it enhance my chance to get better results?